Ike

An AI Answer Engine for GCE OL specification.

Comment

Inspiration

Inspiration struck when faced with the challenge of accessing educational resources tailored to the G.C.E. O/L specification, especially in Sinhala. The scarcity of online materials sparked the idea to develop Ike, an AI-based Answer Engine, to bridge this gap and provide students with a personalized learning experience.

What it does

Ike serves as a customized answer engine for the G.C.E. O/L specification, offering students access to a wealth of educational content extracted from freely available textbooks. Leveraging AI technologies such as vector databases, embeddings, and Langchain, Ike enables users to ask questions in both English and Sinhala, providing relevant answers from the stored text data.

How we built it

We embarked on this journey armed with Python expertise and a thirst for knowledge in AI-related domains. Utilizing vector databases, we stored and indexed the extracted text data for efficient retrieval. Embeddings were employed to represent text in a semantic space, facilitating accurate question-answer matching. Langchain came to the rescue for Sinhala-English translation, ensuring inclusivity for all users.

Challenges we ran into

The road to Ike's creation was not without its hurdles. Integrating various AI technologies seamlessly posed a challenge, requiring meticulous attention to detail and extensive experimentation. Additionally, navigating the complexities of hosting and deployment proved daunting, especially with limited resources and time constraints.Also the task of extracting Sinhala text from both PDFs and images. Converting text from PDFs proved to be relatively straightforward but posed difficulties when dealing with Sinhala characters due to encoding issues. Extracting text from images, especially those containing Sinhala script, presented an even greater challenge, requiring innovative solutions and experimentation with OCR (Optical Character Recognition) technologies, but still couldn't do it properly.

Accomplishments that we're proud of

Despite the challenges encountered, we're proud to have developed Ike, a functional and impactful solution to address the lack of educational resources for the G.C.E. O/L specification. Our achievement lies not only in the technical implementation but also in the potential positive impact Ike can have on students' learning journeys.

What we learned

Through the development of Ike, we gained invaluable insights into the practical applications of AI in education. We deepened our understanding of vector databases, embeddings, and language processing techniques, honing our skills in Python programming along the way. Moreover, we learned the importance of perseverance and adaptability in overcoming obstacles during the development process.

What's next for Ike

The journey doesn't end here for Ike. Moving forward, we aim to refine and expand Ike's capabilities, enriching the dataset with additional educational materials and past papers. Furthermore, we aspire to enhance the user experience by addressing known issues, improving answer accuracy for Sinhala queries, and exploring opportunities for hosting Ike on a suitable platform for wider accessibility. With a commitment to continuous improvement, Ike is poised to become an indispensable tool for students preparing for the G.C.E. O/L examination.

Built With

Share this project:

Updates